Automatic Matching Technology of Two-dimensional Animation Pixels Based on Edge Features

Different 2D pixel animation points have different gray levels, and there will be obvious edge features at the boundary. Using this edge feature, the image can be segmented and matched automatically. It should be noted that the edge feature is not the same as the boundary between objects. The edge feature refers to the mutation of the value of pixels in the image, and the boundary between objects refers to the boundary between objects in the real scene. In real life, there are a variety of two-dimensional animation pixels that can be perceived by us, of which more than 80% of people’s information sources are obtained through the visual system. Vision not only refers to the feeling of light signal, but also includes the whole process of visual information acquisition, transmission, processing, storage and understanding. Both humans and higher animals have developed visual systems, so that things or two-dimensional animation pixels existing in the objective environment can be recognized, which leads to further processing. After the emergence of automatic matching technology, people try to use the camera to obtain two-dimensional animation pixels and convert them into digital signals, and use the computer to realize the whole process of edge feature processing. General machine vision systems work in a certain field, so the target to be identified is in a specific field, and the type is limited, so it can be used as a sample library to accommodate the possible target type template. When judging what the input two-dimensional animation pixels are, you can match and compare the input image with the stored sample library for judgment and recognition. In this paper, a matching method based on edge features using hierarchical matching strategy is proposed. The features such as closed edge feature segments, intersection points and corners of edge features are extracted from continuous edge features as conjugate entities of rough matching. Firstly, the invariant moments of closed edge feature segments, intersection points of edge features and the included angles of edge feature branches at corners are used to determine the features of the same name to be selected; Then these homonymous features to be selected are combined, and the similarity transformation parameters are solved respectively; Rematch 2D animation pixels. The experimental results show that the matching performance is stable and can provide a more accurate initial value for matching.